X · X// 3h agoMODEL RELEASE

Lyria 3 Pro drops for 3-minute tracks

Google DeepMind's Lyria 3 Pro enables the generation of high-fidelity, three-minute songs with explicit structural control over verses, choruses, and bridges. Now available via API in Google AI Studio and to Gemini Advanced subscribers, the model supports multimodal inputs and features SynthID watermarking for professional and creative applications.

// ANALYSIS

Google finally delivers a robust answer to Suno and Udio, prioritizing architectural control and professional quality over simple generation.

–Explicit structural control allows mapping of intros, verses, and bridges, solving long-form musical coherence issues.
–SynthID integration offers a high-standard approach to audio attribution and safety in generative media.
–Multimodal inputs (text, image, video) position the model as a versatile tool for modern content creation workflows.
–API availability in Google AI Studio encourages developers to build custom creative tools atop Google's infrastructure.
–48kHz stereo output and improved vocal clarity significantly narrow the gap between AI generation and studio-quality production.

// TAGS

lyria-3-proaudio-genmultimodalapi

DISCOVERED

3h ago

2026-04-15

PUBLISHED

21d ago

2026-03-25

RELEVANCE

9/ 10

AUTHOR

GoogleDeepMind